Statistictionary | Fuzzy Logic Parlance2020-05-06T15:44:11+05:30

Statistictionary | Fuzzy Logic Parlance

Overview

“Fuzzy”, in one sense, refers to things that aren’t quite clear: fuzzy vision, fuzzy thinking, etc. In computer terms, “fuzzy” is the opposite of true/false Boolean logic; it exists on a continuum between one absolute and another.

We live with the concept of fuzziness every day: Clothes can be wet, dry, or of varying degrees of dampness; colors can be red, purple, or dozens of shades in between; lights can be bright, dim, or almost dark, and so on. So, fuzzy logic is the “something in between” logic that exists between absolute conditions. It’s based on degrees of truth rather than 100% true or 100% false (i.e. the Boolean logic so common in computer science). [1]

 

Fuzzy Logic in Real Life: Washing Machines

Fuzzy logic is leaving the lab and finding many practical applications. Let’s take just one: the washing machine. Washing machines with this feature not only have better performance, they are simple to use, generate less costs, and may even last longer.

What part does fuzzy logic play in the clothes-washing experience? From the Samsung website, we learn that fuzzy logic determines the washing process, including water temperature and levels, agitating time and type, rinse speed and time, and spin speed and time. The fuzzy logic system determines the optimal amount of detergent and water to use (along with many other factors) on each load based on the soil type (i.e. how dirty the clothes are). Optical sensors can even be used to detect the amount of dirt in the water, and a fabric sensor and weight sensor make additional adjustments. [2] A simplified version of this process is illustrated below:

Figure 1: Fuzzy logic in a washing machine, simplified.

Figure 1: Fuzzy logic in a washing machine, simplified.

Interestingly, fuzzy logic mirrors how our brains tend to work, making it essential to the development of AI’s human-like capabilities (e.g. self-learning, problem solving, etc.). This influences how neural networks and other AI applications are utilized.

 

Advantages

The main advantages of fuzzy logic are:

  • It’s often preferable to purely mathematical or symbolic approaches, as most states exist on some kind of continuum.
  • It allows us to measure uncertainty and vagueness .
  • It’s usually simpler, faster to develop, and less computationally demanding than the alternatives.
  • Fuzzy algorithms tend to be more robust and less sensitive to change.

(Adapted from What are the main advantages of applying fuzzy-based decision making?[3])

 

Disadvantages

The main disadvantages of fuzzy logic are:

  • It is often based on assumptions, making the results sometimes inaccurate.
  • Fuzzy systems don’t have the option to use machine learning or neural-network-type pattern recognition.
  • Defining membership functions is difficult, as is setting exact rules.
  • Fuzzy time logic terms can be confused with probability theory terms.
  • The validation and verification of a fuzzy logic system requires extensive testing.

(Adapted from Fuzzy Logic Tutorial: What Is, Application & Example[4].)

 

References

-Authored by Shivli Gupta, Data Scientist at Absolutdata

Technical articles are published from the Absolutdata Labs group, and hail from The Absolutdata Data Science Center of Excellence. These articles also appear in BrainWave, Absolutdata’s quarterly data science digest.

Subscribe to BrainWave